<![CDATA[This study analyzed thematic question items in elementary schools using Iteman software and the Winsteps Rasch Model. The research employed a quantitative method with a descriptive approach, collecting data from 25 multiple-choice questions and 42 student respondents. Reliability scores were 0.641 (classical tests) and 0.68 (modern tests), falling below the desired alpha coefficient of 0.70. Two questions did not meet validity requirements based on MNSQ, ZSTD, and PT Measure. Difficulty levels revealed four very difficult, three too easy, and 18 difficult questions. Classical test analysis showed 10 poor, 12 sufficient, and three very poor discriminating power questions. Modern test analysis identified two students with poor differentiation. Overall, 23 out of 25 questions had well-functioning distractors. Implications include improving measurement quality, curriculum development, student understanding, and learning evaluation at the elementary level.]]>
Copyrights © 2023
